National Repository of Grey Literature 132 records found  1 - 10nextend  jump to record: Search took 0.00 seconds. 
Optimizing Super Mario game tree search
Šosvald, David ; Gemrot, Jakub (advisor) ; Dvořák, Tomáš (referee)
Super Mario Bros. is still actively used as a model game for research in level genera- tion. Every year, the most recent techniques are applied and tested. This lately includes various deep learning and reinforcement learning methods. Many of the level generators use an artificial agent to test levels' playability or to gather playthrough metrics. There- fore, the performance of the level generators is undeniably tied to the performance of the artificial agent used, both in level validation and the computing time needed. In our previous work, we created a new state-of-the-art agent for Super Mario Bros. as a proof of concept when we implemented a more efficient forward model (world simu- lation) for the Mario AI framework. In this work, we continue in that work and focus on optimising how the agents explore the game tree by devising domain-specific heuristics and running extensive parameter searches to tune the agents as much as possible. Thanks to these improvements, a new state-of-the-art agent was created. This new agent should be capable of beating every standard Super Mario Bros. level and it requires less time to solve levels than previous agents. We also present a proof of concept agent that is capable of solving maze-like levels, which is something none of the previous agents was capable of.
Procedural Generator of Short Detective-like Stories
Iudin, Denis ; Gemrot, Jakub (advisor) ; Pešková, Klára (referee)
Procedural generation of interactive stories is still an understudied area, most of the work on which is purely academic, while the application of these technologies in a practical, for a wide audience, area promises good prospects. The reason lies in a large set of factors that complicate the practical use of such systems. In this work, we solve some of them by proposing a new algorithm for this, with the help of which we create a system that generates interactive stories. We visualize the narratives of these stories, and on their basis we create a text quest on the Twine platform.
Group navigation in RTS games using flow networks over flow field regions
Koblížek, Jan ; Gemrot, Jakub (advisor) ; Švancara, Jiří (referee)
This thesis explores the challenges of implementing effective navigation for groups of units in real-time strategy computer games, specifically focusing on the movement of large numbers of homogeneous units across a two-dimensional grid-based map. The thesis presents a pathfinding algorithm that could be used in RTS games. The algorithm enables the units to utilize multiple paths effectively by modeling the unit navigation as a flow network problem. The units use precomputed flow fields during the navigation. This allows for faster pathfinding times by offloading part of the computation to the preprocessing. The algorithm's performance is evaluated against a baseline solution using the A* algorithm and other existing solutions. Comparative analysis will be conducted utilizing maps from the Moving AI 2D Pathfinding Benchmark dataset to assess the efficacy of the proposed solution. 1
Procedural Generator of Short Detective-like Stories
Iudin, Denis ; Gemrot, Jakub (advisor) ; Pešková, Klára (referee)
Procedural generation of interactive stories is still an understudied area, most of the work on which is purely academic, while the application of these technologies in a practical, for a wide audience, area promises good prospects. The reason lies in a large set of factors that complicate the practical use of such systems. In this work, we solve some of them by proposing a new algorithm for this, with the help of which we create a system that generates interactive stories. We visualize the narratives of these stories, and on their basis we create a text quest on the Twine platform.
Analysis of different MCTS implementations of artificial intelligence for the Children of the Galaxy computer game
Guba, Peter ; Gemrot, Jakub (advisor) ; Holeňa, Martin (referee)
Monte Carlo Tree Search (MCTS) is a popular game AI algorithm that searches the state space of a game while using randomized playouts to evaluate new states. There have been many papers published about various adjustments of the original algorithm, however, work that compares multiple of these algorithms together does not seem to exist. This lack of data can make it difficult to decide which variant to use without implementing and testing them which is potentially quite time-consuming. The aim of this thesis is therefore twofold. First to create such a comparison in a specific setting and second to introduce a new variant, WP MCTS, which is based on the idea that one should be able to gather more information from a playout by taking a look at all the states encountered during its computation. For our setting, we chose battles between small armies in a 4X computer game called Children of the Galaxy. The results presented here indicate that many, though not all tested variants outperform basic MCTS in this setting. 1
Sports Manager
Bolemant, Ricardo ; Kofroň, Jan (advisor) ; Gemrot, Jakub (referee)
1 The application is intended for organizers of floorball and hockey competitions, whether league or tournament. It is mainly to help amateur competitions in their management and calculation of various statistics. It simplifies the work for the organizers and saves time thanks to a simple interface. Thanks to this application, even smaller sports competitions will be able to record detailed statistics such as national or international professional competitions. The main goal is to simplify the filling in of match data and their processing.
Cross-platform 2D game framework
Zeman, Tomáš ; Gemrot, Jakub (advisor) ; Holan, Tomáš (referee)
One of the most useful tools for game development is a game framework. It is usually a complex software which offers abstraction of game components such as rendering, physics, sound, user input or AI. The goal of this thesis is to create a simple game framework for 2D games, focused on performance, extensibility and multiplatformity. A second goal of this thesis is implementation of an example game for demonstration of functions and functionality of the framework. Programming language C++ was chosen for development of the framework along with a portion of SDL library. Target platforms were chosen to be Windows and Linux. The example game was successfully implemented and tested on both platforms using most of the framework's capabilities.
Dungeon Master Game for the .NET Platform
Geiger, Petr ; Ježek, Pavel (advisor) ; Gemrot, Jakub (referee)
The goal of this thesis is to reimplement the Dungeon Master game. Currently there exist several clones of this wellknown game. However, compared to them this thesis focuses on aspect stated below. The game is implemented in the C# language using .NET platform. Furthermore, the entire engine is designed towards sustainability and scalability - i. e. that by using this engine it is possible to design slightly different game based on the same principles. Especially, it is easy to add new features to the engine. The engine is also prepared for different input formats of levels. Also the rendering layer of the game engine is completely separate. Due to nature of the project the engine can serve as a representative example of a complex program in programming courses. Powered by TCPDF (www.tcpdf.org)
Navigation Mesh for the Virtual Environment of Unreal Tournament 2004
Tomek, Jakub ; Gemrot, Jakub (advisor) ; Krajíček, Václav (referee)
Navigation mesh is a form of representation of a virtual environment, which allows effective navigation of intelligent virtual agents. This thesis deals with the development of a navigation mesh for the intelligent virtual agents in the environment of the computer game Unreal Tournament 2004. Before the development of the navigation mesh, the navigation graph was the only available representation of the environment that the agents could use for navigation. The navigation graph however provides only limited information about the environment. With the use of the navigation mesh the agents are able to navigate in the environment faster and over shorter paths. Creation of the navigation mesh from the environment's geometry extracted from Unreal Tournament 2004 map files is a part of the thesis. Finally, a new navigation module for Pogamut 3 - a platform that allows creation and controlling of intelligent virtual agents, is implemented. The new navigation module consists of the navigation mesh and a set of navigation algorithms Powered by TCPDF (www.tcpdf.org)
Evaluation of languages for creating intelligent virtual agent behaviors
Vykouk, Ondřej ; Gemrot, Jakub (advisor) ; Bureš, Tomáš (referee)
There are dozens of formalisms specialized programming intelligent virtual agents (IVA). However, there is lack of an evaluation, which would provide the designer of a new agent with an overview of their usability in real-world applications. The goal of this study was evaluation of selected formalisms in order to identify their shortcomings and suggesting improvements. These improvements would allow the designer more accurate control of the IVA's behavior. Evaluation of formalisms was designed based on real-world scenarios and their implementation in different formalisms using the platform Pogamut 3. One of the formalisms is Jason (AgentSpeak language interpreter) and it was connected to that platform as a part of this thesis. During the implementation of the scenarios were identified some problems which make creating the behavior of intelligent virtual agents harder. For these problems was proposed solution in the form of action selection mechanism (ASM) prototype and the basic methodology for designing agents with this ASM. The results of this study could be a starting point for more advanced action selection mechanism, which would allows more precise control of the behavior of intelligent virtual agents. Powered by TCPDF (www.tcpdf.org)

National Repository of Grey Literature : 132 records found   1 - 10nextend  jump to record:
Interested in being notified about new results for this query?
Subscribe to the RSS feed.